For an enterprise with a plurality of branches, in a distributed deployment scenario, generally one cloud storage gateway is deployed in each branch of the enterprise and the plurality of branches shares and uses one cloud service provider. Each cloud storage gateway generally merges a locally stored directory and a cloud storage directory into one directory and presents the merged directory to users in each branch, so that content seen by the users in each branch is inconsistent and directories of the cloud storage gateways need to be synchronized.
Currently, directory synchronization between the cloud storage gateways of the branches is implemented by means of direct communication. Therefore, it is required that the cloud storage gateway of each branch is directly reachable, and for the cloud storage gateway of each branch, a fixed public Internet Protocol (IP) address needs to be assigned, or a data connection is established via a virtual private network (VPN).
However, in an actual deployment environment of a small- to medium-sized enterprise, an address obtained by a cloud storage gateway is usually a dynamic address or even a private address, which makes it difficult to perform direct communication and impossible to perform directory synchronization.